Comparison between sieving and settling-tube determinations of sand sizes by using discriminant analysis

There is much controversy regarding the method that best measures the representative grain sizes used to calculate sediment statistics. Twenty-five sand-sized sediment samples were collected in barrier-island subenvironments and analyzed by a sieving technique and a settling-tube technique. On the basis of graphic sediment statistics as variables, discriminant analysis indicates that the more conventional sieving method correctly characterized 96 percent of the samples by subenvironments, while the settling-tube technique correctly characterized only 72 percent of the samples.